SURE Model as an Advanced Data Science Technique

Abstract

Seemingly unrelated regression equations (SURE) model and it’s associated inferential aspects have been generating substantial applications in various fields such as Statistics, Advanced Econometrics, Data Science Techniques; Business, Management and Marketing fields; physical sciences and Engineering etc. Among the regression based data science techniques, a few data engineers have applied SURE models in their data analysis. Researchers can use SURE techniques as advanced data science techniques in the fields of electrical systems, computer engineering and other areas of engineering and technology. The classical SURE model deals with the sets of linear regression equations by which establishing relationships among the sets of dependent variables and explanatory variables. Several advanced feasible estimation methods exist in the literature used either OLS or GLS residuals in their estimation. In the present research study, due to shortcomings of these residuals, new iterative feasible OLS and feasible GLS estimators have been proposed to estimate the parameters of SURE model with nonspherical first order vector autoregressive errors by using studentized residuals.